Reduced Reproductive Success for a Conditioning Mutant in Experimental Populations of DROSOPHILA MELANOGASTER

Male Drosophila melanogaster that have courted newly-emerged males can modify their subsequent courtship behavior to avoid further courtship with immature males for up to 6 hr (previously reported). Here, it was hypothesized that such an experience-dependent modification would afford a mating advantage to normal males over males that carried a mutation that affects learning and memory. Coisogenic lines were constructed which varied at the dunce gene (dnc+ and dncM14 alleles) in order to test this hypothesis. Whether previously experienced with immature males or not, dnc+ and dncM14 males were indistinguishable in their response and mating efficiency when individually paired with virgin females. However, courtship performance of dnc+ and dncM14 males was different if they were first experienced with immature males and were then individually tested in an artificial population of nine immature males and one virgin female. In this situation, dnc+ males spent much less time in courtship with immature males and achieved copulation in one-third the time required for dncM14 males. As a control, the behavior and mating efficiency of courtship-naive dnc+ and dncM14 males in the artificial population was indistinguishable. In competition for a single virgin female, experienced dncM14 males showed a slight mating advantage over experienced dnc+ males. But when competition by experienced males for a single virgin female took place in the presence of nine immature males, dnc+ males were the successful maters in three-fourths of the trials.     

Synaptic ultrastructure in nerve terminals of Drosophila larvae overexpressing the learning gene dunce

We investigated synaptic ultrastructure of individual nerve ending varicosities at the Drosophila larval neuromuscular junction in transgenic larvae overexpressing the learning gene dunce (dnc) in the nervous system. It was previously shown that cAMP is reduced to one-third normal in these larvae and that they have fewer nerve terminal varicosities and smaller junction potentials, although transmitter release from individual nerve ending varicosities is not significantly altered. We tested the hypothesis that synaptic ultrastructure is modified to compensate for possible reduced efficacy of synaptic transmission resulting from lower than normal cAMP. Synaptic size and number of presynaptic dense bodies (active zone structures) per synapse are modestly enhanced in transgenic larvae overexpressing the dnc gene product and in rutabaga (rut(1)) mutant larvae, which have reduced adenylyl cyclase activity and reduced neural cAMP. The incidence of complex synapses (possessing 2 or more presynaptic dense bodies) was not consistently different in experimental larvae compared to controls. The observations suggest that chronic reduction of cAMP levels in the nervous system of Drosophila larvae, although leading to a modest compensatory change in synaptic structure, does not markedly alter several synaptic ultrastructural parameters which are thought to influence the strength of transmitter release; thus, homeostatic mechanisms do not act to maintain normal-sized junction potentials by altering synaptic structure.     

K(+)-channel blockers restore synaptic plasticity in the neuromuscular junction of dunce, a Drosophila learning and memory mutant.

The effects of K(+)-channel blockers on synaptic transmission in dunce (dnc), a Drosophila learning and memory mutant, were investigated. Larvae dnc mutants lack facilitation and post-tetanic potentiation (PTP) at their motor end-plates; dnc mutants are also deficient in a form of phosphodiesterase, and exhibit abnormally high levels of cyclic adenosine 3',5'-monophosphate (cAMP). A two-microelectrode voltage-clamp was used to record end-plate currents and spontaneous end-plate currents from longitudinal ventrolateral third-instar larval muscle. The K(+)-channel blockers 3,4-diaminopyridine (3,4-DAP) and tetraethylammonium (TEA), at micromolar concentrations, caused a reversible decrease in end-plate current amplitudes both in wild-type and mutant end-plates. In the presence of blockers, a period of high-frequency stimulation (tetanus) of the nerve gave way to a transient increase in the end-plate currents of dnc mutants resembling facilitation and PTP in normal end-plates; 3,4-DAP and TEA also restored facilitation and PTP in normal end-plates after incubation with a non-hydrolysable analogue of cAMP (8Br-cAMP). It is suggested that a specific K+ conductance might be relevant to the lack of synaptic plasticity at the dnc neuromuscular synapses.     

Genetic Analysis of Chromomere 3d4 in DROSOPHILA MELANOGASTER . II. Regulatory Sites for the Dunce Gene

Chromomere 3D4 of the X chromosome of D. melanogaster contains two genes, dunce (dnc) and sperm amotile (sam). Mutations in dnc cause defects in memory formation and female fertility and reduce or eliminate the activity of a cAMP-specific phosphodiesterase designated form II. A fine structure map of this region has been constructed showing the locations of two sam mutations, five dnc mutations and a newly identified locus designated control of fertility (cf) that acts in cis to regulate the female sterility phenotype of dnc. The two sam mutations are separated by 0.02 +/- 0.01 cM, the rightmost being located 0.08 +/- 0.02 cM to the left of the null mutation dncM11. A cluster of null and form II-defective dnc mutations is located 0.04 +/- 0.01 cM to the right of dncM11. The cf locus is 0.06 +/- 0.02 cM to the right of this cluster. The location of the dnc and cf sites identify a region of approximately 0.10 cM that is required for proper expression of dnc+. The dncCK mutation, associated with a reciprocal translocation between 3L and the X, exhibits reduced form II activity and female sterility. This translocation breakpoint has been mapped to the left of the dnc+ gene and is near the breakpoint of Df(1)N64j15 which also reduces expression of dnc+. The effect of these independent chromosomal breaks on the dnc+ gene suggests the existence of a site to the left of dnc+ that is also required for proper expression of the gene.     

The cyclic AMP phosphodiesterase encoded by the Drosophila dunce gene is concentrated in the mushroom body neuropil.

Drosophila dunce (dnc) flies are defective in learning and memory as a result of lesions in the gene that codes for a cAMP-specific phosphodiesterase (PDE). Antibodies to the dnc PDE showed that the most intensely stained regions in the adult brain were the mushroom body neuropil--areas previously implicated in learning and memory. In situ hybridization demonstrated that dnc RNA was enriched in the mushroom body perikarya. The mushroom bodies of third instar larval brains were also stained intensely by the antibody, suggesting that the dnc PDE plays an important role in these neurons throughout their development. The role of the dnc PDE in mushroom body physiology is discussed, and a circuit model describing a possible role of the mushroom bodies in mediating olfactory learning and memory is presented.     

Tetanic stimulation recruits vesicles from reserve pool via a cAMP-mediated process in Drosophila synapses

At Drosophila neuromuscular junctions, there are two synaptic vesicle pools, namely the exo/endo cycling pool (ECP) and the reserve pool (RP). We studied the recruitment process from RP using a fluorescent dye, FMI-43. During high-frequency nerve stimulation, vesicles in RP were recruited for release, and endocytosed vesicles were incorporated into both pools, whereas with low-frequency stimulation, vesicles were incorporated into and released from ECP. Release of vesicles from RP was detected electrophysiologically after emptying vesicles in the ECP of transmitter by a H+ pump inhibitor. Recruitment from RP was depressed by inhibitors of steps in the cAMP/PKA cascade and enhanced by their activators. In rutabaga (rut) with low cAMP levels, mobilization of vesicles from RP during tetanic stimulation was depressed, while it was enhanced in dunce (dnc) with high cAMP levels.     

Characterization of development, behavior and neuromuscular physiology in the phorid fly, Megaselia scalaris

The Phoridae is known as 'scuttle flies' because they walk in rapid bursts of movement with short pauses between. In this study, larval locomotive behavior and development was characterized in the phorid, Megaselia scalaris. Comparison was made with the well-characterized fruit fly model, Drosophila melanogaster. Developmentally, the rate of maturation was consistently slower for Megaselia than Drosophila. This disparity was exaggerated at lower temperatures, particularly during larval development. In addition to slower growth, movements in Megaselia were also slower, as evidenced by reduced rates of larval body wall contractions and mouth hook movements. Megaselia larvae also displayed a unique behavior of swallowing air when exposed to a small pool of liquid. This permitted floating upon immersion and, therefore, might prevent drowning in the natural environment. The anatomical and physiological properties of a neuromuscular junction in the phorid larvae were also examined. The innervation of the motor nerve terminals on the ventral abdominal muscle (m6) is innervated by Type Ib and Is axons, similar to Drosophila. As in Drosophila, the Is terminals produce larger excitatory postsynaptic potentials (EPSPs) than the Ib. The amplitudes of the EPSPs in M. scalaris were reduced compared to those of D. melanogaster, but unlike D. melanogaster the EPSPs showed marked facilitation when stimulated with a 20 Hz train. We conclude that there may be differences in synaptic structure of the nerve terminals that could account for the different electrophysiological behaviors.     

Modulation of dihydropyridine-sensitive calcium channels in Drosophila by a cAMP-mediated pathway.

Drosophila has proved to be a valuable system for studying the structure and function of ion channels. However, relatively little is known about the regulation of ion channels, particularly that of Ca2+ channels, in Drosophila. Physiological and pharmacological differences between invertebrate and mammalian L-type Ca2+ channels raise questions on the extent of conservation of Ca2+ channel modulatory pathways. We have examined the role of cyclic adenosine monophosphate (cAMP) cascade in modulating the dihydropyridine (DHP)-sensitive Ca2+ channels in the larval muscles of Drosophila, using mutations and drugs that disrupt specific steps in this pathway. The L-type (DHP-sensitive) Ca2+ channel current was increased in the dunce mutants, which have high cAMP concentration owing to cAMP-specific phosphodiesterase (PDE) disruption. The current was decreased in the rutabaga mutants, where adenylyl cyclase (AC) activity is altered thereby decreasing the cAMP concentration. The dunce effect was mimicked by 8-Br-cAMP, a cAMP analog, and IBMX, a PDE inhibitor. The rutabaga effect was rescued by forskolin, an AC activator. H-89, an inhibitor of protein kinase-A (PKA), reduced the current and inhibited the effect of 8-Br-cAMP. The data suggest modulation of L-type Ca2+ channels of Drosophila via a cAMP-PKA mediated pathway. While there are differences in L-type channels, as well as in components of cAMP cascade, between Drosophila and vertebrates, main features of the modulatory pathway have been conserved. The data also raise questions on the likely role of DHP-sensitive Ca2+ channel modulation in synaptic plasticity, and learning and memory, processes disrupted by the dnc and the rut mutations.     

Population density regulates Drosophila synaptic morphology in a Fasciclin-II-dependent manner

Genetic analysis of the Drosophila larval neuromuscular junction has identified some of the key molecules that regulate synaptic plasticity. Among these molecules, the expression level of Fasciclin II (FasII), a homophilic cell adhesion molecule, is critically important for determining the final form of the neuromuscular junction. Genetic reduction of FasII expression by 50% yields more elaborate nerve terminals, while a greater reduction in expression, to 10% of wild-type, yields a substantial reduction in the nerve terminal morphology. Importantly, regulation of FasII expression seems to be the final output for several genetic manipulations that transform NMJ morphology. In an effort to understand the importance of this regulatory pathway in the normal animal, we have undertaken studies to identify environmental cues that might be important for initiating FasII-dependent changes in synaptic plasticity. Here we report on the relationship between larval population density and synaptic morphology, synaptic strength, and FasII levels. We raised Drosophila larvae under conditions of increasing population density and found an inverse exponential relationship between population density and the number of synaptic boutons, the number of branches, and the length of branches. We also observed population-dependent alteration in FasII levels, with lower densities having less FasII at the synapse. The correlation between density and morphological change was abrogated in larvae constitutively expressing FasII, and in wild-type larvae grown on soft culture medium. Together these data show that environmental cues can induce regulation of FasII. Interestingly, however, the quantal content of synaptic transmission was not different among the different population densities, suggesting that other factors contribute to maintaining synaptic strength at a defined level.     

Retrograde Gbb signaling through the Bmp type 2 receptor wishful thinking regulates systemic FMRFa expression in Drosophila

Amidated neuropeptides of the FMRFamide class regulate numerous physiological processes including synaptic efficacy at the Drosophila neuromuscular junction (NMJ). We demonstrate here that mutations in wishful thinking (wit) a gene encoding a Drosophila Bmp type 2 receptor that is required for proper neurotransmitter release at the neuromuscular junction, also eliminates expression of FMRFa in that subset of neuroendocrine cells (Tv neurons) which provide the systemic supply of FMRFa peptides. We show that Gbb, a Bmp ligand expressed in the neurohemal organ provides a retrograde signal that helps specify the peptidergic phenotype of the Tv neurons. Finally, we show that supplying FMRFa in neurosecretory cells partially rescues the wit lethal phenotype without rescuing the primary morphological or electrophysiological defects of wit mutants. We propose that Wit and Gbb globally regulate NMJ function by controlling both the growth and transmitter release properties of the synapse as well as the expression of systemic modulators of NMJ synaptic activity.     

Cellular bases of activity-dependent paralysis in Drosophila stress-sensitive mutants

Stress-sensitive mutants in Drosophila have been shown to exhibit activity-dependent defects in neurotransmission. Using the neuromuscular junction (NMJ), this study investigates synaptic function more specifically in two stress-sensitive mutants: stress-sensitive B (sesB), which encodes a mitochondrial ADP/ATP translocase (ANT); and Atpalpha(2206), a conditional mutant of the Na+/K+ ATPase alpha-subunit. Mechanical shock induces a period of brief paralysis in both homozygous and double heterozygous mutants, but further analysis revealed distinct activity-dependent neurotransmission lesions in each mutant. Basal neurotransmission appeared similar to wild-type controls in both mutants under low frequency stimulation. High frequency stimulation, however, caused pronounced synaptic fatigue as well as slow and incomplete synaptic recovery in sesB mutants while Atpalpha(2206) mutants displayed an increase (25-fold) in synaptic failures. Perhaps to compensate for these activity dependent defects, the neuromuscular synapse was found to be overgrown in both mutants. Passive electrotonic stimulation, which initiates synaptic transmission independent of action potentials, ameliorated synaptic failures and resulted in increased neurotransmission amplitude in Atpalpha(2206) mutants. In addition, spontaneous synaptic vesicle fusion rates were increased in Atpalpha(2206) mutants, suggesting that, in the absence of action potential requirements, these synaptic terminals are healthy, if not hyperactive. Dye labeling studies revealed aberrant synaptic vesicle cycling in sesB mutants indicating a reduction of functional synaptic vesicles. We therefore postulate that both stress-sensitive mutants harbor unique neurotransmission defects: Atpalpha(2206) mutants are unable to maintain ionic gradients required during repetitive action potential propagation, and sesB mutants cannot maintain synaptic vesicle cycling during periods of high demand.     

Signaling for Vesicle Mobilization and Synaptic Plasticity

The hypothesis that release of classical neurotransmitters and neuropeptides is facilitated by increasing the mobility of small synaptic vesicles (SSVs) and dense core vesicles (DCVs) could not be tested until the advent of methods for visualizing these secretory vesicles in living nerve terminals. In fact, fluorescence imaging studies have only since 2005 established that activity increases secretory vesicle mobility in motoneuron terminals and chromaffin cells. Mobilization of DCVs and SSVs appears to be due to liberation of hindered vesicles to promote quicker diffusion. However, F-actin and synapsin, which have been featured in mobilization models, are not required for activity-dependent increases in the mobility of DCVs or SSVs. Most recently, the signaling required for sustained mobilization has been identified for Drosophila motoneuron DCVs and shown to increase synaptic transmission. Specifically, presynaptic endoplasmic reticulum ryanodine receptor-mediated Ca2+ release activates Ca2+/calmodulin-dependent kinase II to mobilize DCVs and induce post-tetanic potentiation (PTP) of neuropeptide release in the Drosophila neuromuscular junction. The shared signaling for increasing vesicle mobility and PTP links vesicle mobilization and synaptic plasticity.     

Evidence for postsynaptic modulation of muscle contraction by a Drosophila neuropeptide

DPKQDFMRFamide, the most abundant FMRFamide-like peptide in Drosophila melanogaster, has been shown previously to enhance contractions of larval body wall muscles elicited by nerve stimulation and to increase excitatory junction potentials (EJPs). The present work investigated the possibility that this peptide can also stimulate muscle contraction by a direct action on muscle fibers. DPKQDFMRFamide induced slow contractions and increased tonus in body wall muscles of Drosophila larvae from which the central nervous system had been removed. The threshold for this effect was approximately 10(-8)M. The increase in tonus persisted in the presence of 7x10(-3)M glutamate, which desensitized postsynaptic glutamate receptors. Thus, the effect on tonus could not be explained by enhanced release of glutamate from synaptic terminals and, thus, may represent a postsynaptic effect. The effect on tonus was abolished in calcium-free saline and by treatment with L-type calcium channel blockers, nifedipine and nicardipine, but not by T-type blockers, amiloride and flunarizine. The present results provide evidence that this Drosophila peptide can act postsynaptically in addition to its apparent presynaptic effects, and that the postsynaptic effect requires influx through L-type calcium channels.     

Anti-agrin staining is absent at abandoned synaptic sites of frog neuromuscular junctions

The neuromuscular junction is a plastic structure and is constantly undergoing changes as the nerve terminals that innervate the muscle fiber extend and retract their processes. In vivo observations on developing mouse neuromuscular junctions revealed that prior to the retraction of a nerve terminal the acetylcholine receptors (AChRs) under that nerve terminal disperse. Agrin is a protein released by nerve terminals that binds to synaptic basal lamina and directs the aggregation of AChRs and acetylcholinesterase (AChE) in and on the surface of the myotube. Thus, if the AChRs under a nerve terminal disperse, then the cellular signaling mechanism by which agrin maintains the aggregation of those AChRs must have been disrupted. Two possibilities that could lead to the disruption of the agrin induced aggregation are that agrin is present at the synaptic basal lamina but is unable to direct the aggregation of AChRs, or that agrin has been removed from the synaptic basal lamina. Thus, if agrin were blocked, one would expect to see anti-agrin staining at abandoned synaptic sites; whereas if agrin were removed, anti-agrin staining would be absent at abandoned synaptic sites. We find that anti-agrin staining and α-bungarotoxin staining are absent at abandoned synaptic sites. Further, in vivo observations of retracting nerve terminals confirm that agrin is removed from the synaptic basal lamina within 7 days. Thus, while agrin will remain bound to synaptic basal lamina for months following denervation, it is removed within days following synaptic retraction. © 1996 John Wiley & Sons, Inc.     

Drosophila larval neuromuscular junction's responses to reduction of cAMP in the nervous system

We investigated the effects of chronically lowered cyclic adenosine monophosphate (cAMP) on the morphology and physiology of the Drosophila larval neuromuscular junction, using two fly lines in which cAMP was significantly lower than normal in the nervous system: (a) transgenic flies in which the dunce (dnc) gene product was overexpressed in the nervous system, and (b) flies mutant for the rutabaga gene (rut¹) which have reduced adenylyl cyclase activity. In comparison with controls, larvae with reduced cAMP exhibited a smaller number of synaptic varicosities. This effect was more pronounced in transgenic larvae, in which the reduction of neural cAMP was more pronounced. Synaptic transmission was also reduced in both cases, as evidenced by smaller excitatory junctional potentials (EJPs). Synaptic currents recorded from individual synaptic varicosities of the neuromuscular junction indicated almost normal transmitter release properties in transgenic larvae and a modest impairment in rut¹ larvae. Thus, reduction in EJP amplitude in transgenic larvae is primarily due to reduced innervation, while in rut¹ larvae it is attributable to the combined effects of reduced innervation and a mild impairment of transmitter release. We conclude that the major effect of chronically lowered cAMP is reduction of innervation rather than impairment of transmitter release properties. © 1999 John Wiley & Sons, Inc. J Neurobiol 40: 1–13, 1999     

Analysis of adhesion molecules and basement membrane contributions to synaptic adhesion at the Drosophila embryonic NMJ

Synapse formation and maintenance crucially underlie brain function in health and disease. Both processes are believed to depend on cell adhesion molecules (CAMs). Many different classes of CAMs localise to synapses, including cadherins, protocadherins, neuroligins, neurexins, integrins, and immunoglobulin adhesion proteins, and further contributions come from the extracellular matrix and its receptors. Most of these factors have been scrutinised by loss-of-function analyses in animal models. However, which adhesion factors establish the essential physical links across synaptic clefts and allow the assembly of synaptic machineries at the contact site in vivo is still unclear. To investigate these key questions, we have used the neuromuscular junction (NMJ) of Drosophila embryos as a genetically amenable model synapse. Our ultrastructural analyses of NMJs lacking different classes of CAMs revealed that loss of all neurexins, all classical cadherins or all glutamate receptors, as well as combinations between these or with a Laminin deficiency, failed to reveal structural phenotypes. These results are compatible with a view that these CAMs might have no structural role at this model synapse. However, we consider it far more likely that they operate in a redundant or well buffered context. We propose a model based on a multi-adaptor principle to explain this phenomenon. Furthermore, we report a new CAM-independent adhesion mechanism that involves the basement membranes (BM) covering neuromuscular terminals. Thus, motorneuronal terminals show strong partial detachment of the junction when BM-to-cell surface attachment is impaired by removing Laminin A, or when BMs lose their structural integrity upon loss of type IV collagens. We conclude that BMs are essential to tie embryonic motorneuronal terminals to the muscle surface, lending CAM-independent structural support to their adhesion. Therefore, future developmental studies of these synaptic junctions in Drosophila need to consider the important contribution made by BM-dependent mechanisms, in addition to CAM-dependent adhesion.     

Presynaptic effectors contributing to cAMP-induced synaptic potentiation in Drosophila

cAMP analogs and activation of adenylyl cyclase by forskolin strongly potentiate synaptic transmission at the Drosophila neuromuscular junction. These effects are generally attributed to activation of cAMP-dependent protein kinase. Recent reports on crustacean and mammalian synapses have implicated other cAMP-dependent effectors in synaptic potentiation. Drosophila neuromuscular junctions were tested for effects of two known cAMP-dependent effectors: hyperpolarization-activated, cyclic nucleotide-regulated channels (HCNCs) and guanine nucleotide exchange protein activated by cAMP (Epac). Forskolin-induced enhancement of synaptic transmission was drastically reduced by a blocker of HCNCs, but not completely eliminated. A specific agonist for Epac modestly enhanced synaptic potentials. This agonist also stabilized their amplitudes in the presence of a blocker of HCNCs. The observations implicate HCNCs and Epac in cAMP-dependent potentiation that does not require cAMP-dependent protein kinase, indicating that additional previously unexplored factors contribute to synaptic plasticity in Drosophila. Genetic and molecular techniques available for Drosophila can be used to define the underlying molecular basis for cAMP-dependent synaptic potentiation.     

Investigating the safety factor at an invertebrate neuromuscular junction

Fidelity of synaptic transmission is essential at the neuromuscular junction (NMJ). To ensure that transmission does not fail, vertebrate motoneurons often release more neurotransmitter than is required for muscle contraction. This safety factor allows some loss of synaptic function without failure of muscle contraction. It is not known whether a similar mechanism operates at the invertebrate neuromuscular junction. In our study of the Drosophila NMJ, we find that glutamate receptor mutants can exhibit a substantial decrease in synaptic function while maintaining muscle contraction. The persistence of neuromuscular function in these mutants is not explained by synaptic facilitation, temporal summation of high frequency stimuli, or a hyperpolarizing shift in the activation range of muscle calcium channels. Instead, the attenuated synaptic response is sufficient to drive muscle contraction. Quantitative analysis of the decrease in synaptic transmission in these mutants implies that at the wild-type NMJ there is an approximately five- to ninefold excess in released transmitter. Hence, the presence of a synaptic safety factor is a conserved feature of neuromuscular organization in both invertebrates and vertebrates.     

Cullin-5 and cullin-2 play a role in the development of neuromuscular junction and the female germ line of <Emphasis Type="Italic">Drosophila</Emphasis>

Cullins confer substrate specificity to E3-ligases which are multi-protein complexes involved in ubiquitin-mediated protein degradation or modification. There are six cullin genes in Drosophila melanogaster. We have raised an antibody against Cul-5 and demonstrated that it expresses in neuronal and non-neuronal cells throughout development. In the embryonic tracheal system, Cul-5 is enriched at fusion sites together with E-Cadherin and Fasciclin III. Mutations of cul-5 do not affect tracheal development but do show defects in the organization of synaptic boutons at the larval neuromuscular junction where the protein is expressed in a subset of motoneuron terminals. Loss of function of another cullin gene ‘cul-2’ results in similar defects at the larval neuromuscular junction although cul-2;cul-5 double mutants do not show an enhanced phenotype. Both cul-2 and cul-5 mutants show similar aberrations in the development of female germ line. Our results suggest that both of these cullin proteins participate in similar developmental processes.     

Morphological differences between excitatory and inhibitory nerve terminals in cockroach coxal muscles

Two morphologically distinct types of neuromuscular junction on the coxal leg muscles of the cockroach, Periplaneta americana, which have been physiologically described as innervated by fast, slow and inhibitory nerve fibers, have been found. In one type of neuromuscular junction the axon terminal contains many round clear synaptic vesicles and contacts several sarcoplasmic extensions from the muscle fiber. The muscle processes adhere to the axon terminal for a short distance (short contact or SC type). The axon terminal of the other type of neuromuscular junction directly contacts the muscle fiber and no extensions of the muscle fiber are formed. The contact region is comparatively long (long contact or LC type). The nerve terminal contains many polymorphic synaptic vesicles. From a correlation of the present morphological findings and the previous physiological results, it may be suggested that the SC type of nerve terminal represents both fast and slow nerve terminals and the inhibitory terminal is of the LC type.